English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Unsupervised domain adaptation (UDA) aims to learn transferable knowledgefrom a labeled source domain and adapts a trained model to an unlabeled targetdomain. To bridge the gap between source and target domains, one prevailingstrategy is to minimize the distribution discrepancy by aligning their semanticfeatures extracted by deep models. The existing alignment-based methods mainlyfocus on reducing domain divergence in the same model layer. However, the samelevel of semantic information could distribute across model layers due to thedomain shifts. To further boost model adaptation performance, we propose anovel method called Attention-based Cross-layer Domain Alignment (ACDA), whichcaptures the semantic relationship between the source and target domains acrossmodel layers and calibrates each level of semantic information automaticallythrough a dynamic attention mechanism. An elaborate attention mechanism isdesigned to reweight each cross-layer pair based on their semantic similarityfor precise domain alignment, effectively matching each level of semanticinformation during model adaptation. Extensive experiments on multiplebenchmark datasets consistently show that the proposed method ACDA yieldsstate-of-the-art performance.

Attention-based Cross-Layer Domain Alignment for Unsupervised Domain Adaptation